test_model.py

# -*- coding: utf-8 -*-
"""
Created on Tue Oct  2 22:16:06 2018

@author: Trushant Kalyanpur
"""

import numpy as np
import matplotlib.pyplot as plt
import pickle
from regression_model import regression_model
from sklearn.metrics import mean_squared_error,r2_score,mean_absolute_error
import pandas as pd

#import xgboost as xgb

import ml_functions as dp

#Import test data
test_df  = pd.read_csv('new_test.csv')


#Initialize  and load pre-trained model
regressor = regression_model()
model_path = 'random_forest_ww46'
regressor.loadModel(model_path)

#Impute missing values with mean
#X_imputed = dp.impute_missing_mean(test_df)
#X = X_imputed

#Drop sale price since this is the target value
X_test = test_df.drop("SalePrice",axis=1)



#Target feature
y_true = test_df["SalePrice"]


#Predicted values 

#uncomment for non-xgboost models
#y_pred = regressor.predict(X_test)

#Uncomment for xgboost
y_pred = regressor.predict(X_test.values)

#Calculate Mean Absolute Error and R2 score based on predictions
mae = mean_absolute_error(y_true,y_pred)
r2 = r2_score(y_true,y_pred)

print ("Mean absolute error=",mae)
print ("R2 score=",r2)